What is ML model governance?
Machine learning model governance
Since machine learning is a relatively new discipline, there are many inefficiencies that need to be addressed in ML processes and models. Since the field is so nascent there is also a lack of standards in different aspects of ML processes and models. Strong model governance and management can solve many of these issues. Many companies are still implementing different open source models optimized using different techniques to achieve results. Governance helps to provide a scaffolding of sorts for important parts of the ML process to extract better ROI.
Model governance
Machine learning model governance is the framework for an organization to control access, implement policy, and track activity for their models.
This includes setting the rules and controls for machine learning models in production, such as access control, testing, validation, the monitoring of model results, and effective documentation and versioning of models.
Continuous monitoring for ML models is considered to be the first and perhaps the most important step in model governance. It requires thinking from a dataset as well as a model perspective as explained in the article. On a more ML Ops level, it is important to have version control across your ML…
